QDevelop, a new IDE for Qt4

[elcuco]

trying to init completion on a null class name

Before revision 200 it would crahs your application. I made that a qDebug to see how much my "fix" holds up in the real world. Happy it does.

On your setup it failed to load some tree classes, since your project contains some features not supported by the qmake parser. We should discuss about the handling of this issue.

That said, you tested trunk, which is still been developed (did you see that there is no option to disable the white space markings...? I am still evaluating this feature).

Can you edit your previous message and put numbers instead of bullets?

[marcel]

OK. This has been fun enough...
But why in the world you people work on separate IDE's and not join forces and create a single and better ( and bigger ) IDE?

I am sorry to say this, but none of the Qt open source IDEs( maybe except KDevelop) cannot compare to commercial ones or Eclipse. And this is due to missing functionality...

I agree that competition is good, but all that was fun went away a long time ago...
I am sure you can gather cool things from all the IDEs and build one tough motherf***er.

This goes to fullmetalcoder, elcuco, magland and everybody that has a homemade IDE.

EDIT: I forgot to add a smiley:

Regards

[fullmetalcoder]

On your setup it failed to load some tree classes, since your project contains some features not supported by the qmake parser. We should discuss about the handling of this issue.

This does not come from edyuk.pro AFAIK things are a wee bit weirder because I got these error message only once (not once for all time but at a single launch of QDevelop)... I have not been able to reproduce it...

Can you edit your previous message and put numbers instead of bullets?

I guess I could if I were still allowed to edit it... But even then I doubt I'd do it because it's not up to me to prioritize the items there...

But why in the world you people work on separate IDE's and not join forces and create a single and better ( and bigger ) IDE?

Man this has been asked a hundred of times (at the very least)... Because we've all started something which we are proud of and we are reluctant to abandon it, some more than the others... Some other nagging questions would be : How do we name it? Who's gonna lead the team? Why should we restart from scratch now that things are at last working? But then which code should we take as a basis? etc...
If you are able to answer these questions please do and you may be able to convince all the people involved (i.e. those you mentioned + jlbrd and Pasnox, who would probably be the hardest to convince BTW...)

[sunil.thaha]

I support marcel's suggestion
Its always better to work in a team when the you have a common goal.
Most of the IDE's cannot be compared to that of KDevelop
And I believe its not because you guys are technically incompetent.

Man this has been asked a hundred of times (at the very least)... Because we've all started something which we are proud of and we are reluctant to abandon it, some more than the others...

In short ... Can I call it "Ego" , which says "If my design is not good, I will not accept that yours is better "

Some other nagging questions would be : How do we name it? Who's gonna lead the team?

Does that really matter when everyone is working for a common goal?

Why should we restart from scratch now that things are at last working? But then which code should we take as a basis? etc...

You might need to, but the lessons learned from your projects are not lost? I am sure it will be put to a good use; only if are willing to accept others views also

If you are able to answer these questions please do and you may be able to convince all the people involved (i.e. those you mentioned + jlbrd and Pasnox, who would probably be the hardest to convince BTW...)

So does that mean you are willing to join someone for common goal ?

[fullmetalcoder]

Most of the IDE's cannot be compared to that of KDevelop

Well, AFAIK KDevelop sucks when it comes to develop Qt 4 applications. It surely has many great features and a huge community but :

• no Qt 4 designer integration
• Qt 4 completion doesn't work (at least I did not manage to make it work... Am I stupid or is it a bug?)
• Qt 4 project management is known to be incomplete and/or buggy (did not try the very latest code though...)

Of course we cannot be compared to KDevelop due to its modularity and huge community but, in our domain, we can outrank it...

In short ... Can I call it "Ego" , which says "If my design is not good, I will not accept that yours is better "

Does that really matter when everyone is working for a common goal?

It may matter for some people... Because of the "ego"...

You might need to, but the lessons learned from your projects are not lost? I am sure it will be put to a good use; only if are willing to accept others views also

The lessons I see so far :

1. We need plugins. I've been the first to switch the design and both Pasnox and jlbrd followed me later on
2. We need to agree about the UI : static or dynamic?
3. We need to agree about the editing layer. QDevelop and HiQt still uses QTextEdit under the hoods, hence higher memory consumption, slower loading and highlighting and possible lockups with large files. On the other hand Monkey 2 will use QScintilla and I rewrote a text editing from scratch (well, from QtCore and QPainter actually...)

So does that mean you are willing to join someone for common goal ?

Yep. I'm open to talks and if we can find an agreement on where to start and where to go and how we plan to get there I'll be willing to do it. However I really doubt we'll manage to agree because Pasnox has been working hard on Monkey 2 (we're talking a bunch) which he plans to release within a couple of monthes, because until now QDevelop team clung to its code/design/..., because magland has started is project knowing about the others, instead of joining one, ...
That's quite sad because I'd really like such a thing to happen since I could at last do what I like and master most (hard work on internals, e.g. completion backend, project parsers, HL engine, ... ) without having to spend all my time on usability matters that I really don't enjoy (hey! nobody is perfect...)

[magland]

Advantage of open source is that we collaborate without being on the same team! A successful IDE will be one that can take the best ideas from all IDEs and tie them together on top of a solid foundation - IMHO has not been achieved yet.

For example, if the thing is coded right, the editor component should be completely independent of the rest of the IDE. I've already announced that the next release of Handcoder's IDE will be able to switch seamlessly between QTextEdit and QScintilla for the editor. AND as soon as fmc's QCodeEdit is ready, I plan to just plug it in. That's efficient collaboration I say. If components are not modular enough to be plugged into another application without too much effort - then they are not worth using anyway. I have borrowed syntax highlighting and some icons from QDevelop (who probably derived them from elsewhere).

Handcoder's IDE is fundamentally different than other IDEs and serves a different purpose - one thing, it treats the .pro file as source code.

[fullmetalcoder]

For example, if the thing is coded right, the editor component should be completely independent of the rest of the IDE. I've already announced that the next release of Handcoder's IDE will be able to switch seamlessly between QTextEdit and QScintilla for the editor. AND as soon as fmc's QCodeEdit is ready, I plan to just plug it in. That's efficient collaboration I say.

Where do you see the collaboration here? Don't get me wrong, you're absolutely free to do that and it could reveal extremely cool though it will probably require a bunch of work from you to create an abstraction layer over 3 very different frameworks (in terms of API and impl I mean). The fact is that you just take things ONCE they are ready, which is normal to make your IDE stable, but you DO NOT cooperate allow these modules to get ready sooner... That would be collaboration.

If components are not modular enough to be plugged into another application without too much effort - then they are not worth using anyway.

True enough. That's basically what I'm trying to do with editing but also completion/class browsing backend and a couple other stuff and I'd say I did it quite well so far...

I have borrowed syntax highlighting and some icons from QDevelop (who probably derived them from elsewhere).

HL code of QDevelop comes from QSA AFAIK (and there should be a notice left to respect GPL...

Handcoder's IDE is fundamentally different than other IDEs and serves a different purpose - one thing, it treats the .pro file as source code.

Your approach of project handling is not such a fundamental difference... The fundamentals of an IDE are :

• loading project files
• editing sources
• compiling

All the rest is good to have but not "fundamental", at least that's my understanding...

[KaptainKarl]

I think most users are altering the expectation of what an IDE is. Certainly most expect there to be an integrated debugger along with projects, code editing, and compiling.

Karl

[marcel]

I think most users are altering the expectation of what an IDE is. Certainly most expect there to be an integrated debugger along with projects, code editing, and compiling.

Karl

I, for one, don't like to waste too much time/thinking on debugging. Just want to debug, get the info I need, and get on with solving the real problems.

How much time does one waste with command line debuggers?

And of course, there's that little thing with information being presented to you visually.

Regards

[jlbrd]

In this thread I would like to see posts about QDevelop. Please create an new thread for your general discussions about IDE and other.

Thank you

[Morea]

Is the code completion working? If yes, how to notice it?
I started writing

class slask{
public:
int foo();
int xabcdef;

};

in a h file, and in the cpp file that included the above h file, I wrote

#include "slask.h"
int slask::foo()
{

xabc AND HERE WERE NO CODE COMPLETION!!!!!!
}

[jlbrd]

Hi,
The completion works now with the words alone as you want (since 10 minutes in fact). You must build the last trunk present in the svn.

[eekhoorn12]

Hey
I am also trying this program but when i try to compile i get the following:

Compileren (make)...
C:/Qt/4.3.0/bin/mingw32-make.exe -f Makefile.Debug all
mingw32-make.exe[1]: Entering directory `C:/Qt/4.3.0/Projects/gotocell'
g++ -c -g -frtti -fexceptions -mthreads -Wall -DUNICODE -DQT_LARGEFILE_SUPPORT -DQT_DLL -DQT_GUI_LIB -DQT_CORE_LIB -DQT_THREAD_SUPPORT -DQT_NEEDS_QMAIN -I"..\..\include\QtCore" -I"..\..\include\QtCore" -I"..\..\include\QtGui" -I"..\..\include\QtGui" -I"..\..\include" -I"c:\MySql\include" -I"." -I"..\..\include\ActiveQt" -I"tmp\moc\debug_shared" -I"." -I"..\..\mkspecs\win32-g++" -o tmp\obj\debug_shared\gotocelldialog.o gotocelldialog.cpp
mingw32-make.exe[1]: Leaving directory `C:/Qt/4.3.0/Projects/gotocell'
g++ wordt niet herkend als een interne
of externe opdracht, programma of batchbestand.
mingw32-make.exe[1]: *** [tmp\obj\debug_shared\gotocelldialog.o] Error 1
C:\Qt\4.3.0\bin\mingw32-make.exe: *** [debug-all] Error 2

Compileren (make)...
C:/Qt/4.3.0/bin/mingw32-make.exe -f Makefile.Debug all
mingw32-make.exe[1]: Entering directory `C:/Qt/4.3.0/Projects/gotocell'
g++ -c -g -frtti -fexceptions -mthreads -Wall -DUNICODE -DQT_LARGEFILE_SUPPORT -DQT_DLL -DQT_GUI_LIB -DQT_CORE_LIB -DQT_THREAD_SUPPORT -DQT_NEEDS_QMAIN -I"..\..\include\QtCore" -I"..\..\include\QtCore" -I"..\..\include\QtGui" -I"..\..\include\QtGui" -I"..\..\include" -I"c:\MySql\include" -I"." -I"..\..\include\ActiveQt" -I"tmp\moc\debug_shared" -I"." -I"..\..\mkspecs\win32-g++" -o tmp\obj\debug_shared\gotocelldialog.o gotocelldialog.cpp
mingw32-make.exe[1]: Leaving directory `C:/Qt/4.3.0/Projects/gotocell'
g++ wordt niet herkend als een interne
of externe opdracht, programma of batchbestand.
mingw32-make.exe[1]: *** [tmp\obj\debug_shared\gotocelldialog.o] Error 1
C:\Qt\4.3.0\bin\mingw32-make.exe: *** [debug-all] Error 2

To copy to clipboard, switch view to plain text mode 

g++ wordt niet herkend als een interne
of externe opdracht, programma of batchbestand.
means 'g++' is not recognized as an internal or external command

all the external programs except gdb and ctag are in C:/Qt/4.3.0/bin/ zo i set the location of those programs to that directory.

[jlbrd]

Mingw is not properly installed, it is required to re-install it. I see in your results that make is called in the Qt/bin directory, certainly not a good idea.

[eekhoorn12]

Reinstalled it and now it works thanks

[GreyGeek]

QDevelop is great!

Not too fat, not to skinny. Just about right. If I wanted the kitchen sink I'd use emacs or Eclipse. If I wanted code completion to add properties and methods that didn't apply to the object I'm using I'd use Visual Studio.

I like the way QDevelop works. I like the size of its footprint. it's easy to use and it's fast.

Polish it, but don't turn it into a pig.

[jlbrd]

I work always to improve qdevelop. But also to keep it easy to use. It's for this reason that Designer is not embedded in the IDE and Assistant too. I work on the completion for the next version. It will add the members and properties of the inherited classes (sorry). The completion will be also very faster than the lasted versions. But although these modifications, I do not want to change it into a pig or other animal.

Ps: Your expressions (kitchen, pig etc.) are funny when they are translated in French . Yes Google translate is my friend to write and read on this forum.

Thank you to appreciate and use qdevelop.

[brokensword]

Hi. Let me insert my 5 coins. Yes, QDevelop is good. I've tried to use it professionally. Everything any real developer needs is done on 5+. Except one, most important thing - debugging abilities. They are on 2-. If you put "more" debugging support to QDevelop - it will became hundred times popularer. There are no use of QDevelop for developer without debugging, because probably 90% of all time developer debugs his code.
There are no any good gdb front-ends on Windows platforms (I would use it in tandem with QDeveloper if it existed).

Believe me - code completion is also important, but it's nothing compared to debugging. Concentrate on debugging support and you will beat MS VS

[jlbrd]

But qdevelop provide debugging functions.
Install gdb and set his location in menu "Tools|External tools", change the properties of the project to "debug" in "project properties", put breakpoints by right-clicks in line numbers (or F9, or menu), rebuild all, launch "Debug|Debug" or F5, the program stops on breakpoints. After use the menu "Debug" or the Toolbar.

[brokensword]

Existing functions are usefull just for simplest bugs fixing. For full program developing and supporting cycle following features are needed:

- View disassembled code (with tracing ability)
- Inserting breakpoints (memory r\w\e + conditional)
- Memory dump view (with drag/drop from disassembled code window etc)
- CPU Registers view\change

For example, you can look at debugging features of SlickEdit or Eclipse. Everything regarding debugging is not just nice features of IDE - they are absolutely MUST to have to fix most of bugs of usual code.

Until QDevelop haven't at least listed above features - it's usefull just for amateuring development. After their implementation it becames 1000 times powerfull as IDE for developing then now and for me it becames powerest then MS VS.